From what you’ve seen in the field, why do data warehouse projects fail?
For the past 16 years I’ve been involved in the data warehousing / business intelligence industry. I’ve seen an amazing amount of progress and yet, nearly every day I meet with a potential customer that has fallen victim to one or more of the major pitfalls inherent to data warehouse initiatives. Here are some tips to avoid common mistakes.
- Don’t try to boil the ocean! I recently met with a company that designed what they called a “supertable” to report off of as their first phase. This is a recipe for disaster for numerous reasons – performance issues, ease of reporting, and delayed project timelines just to name a few. Set the expectation with the business that quick wins are essential in the first phase of the data warehouse project so that risk and timelines are minimized.
- Hire the right IT resources with the correct skills and background. This same company decided to utilize employees from within one of their division’s IT departments with zero experience in data warehousing or business intelligence. Just because someone has regression testing or systems design experience, it doesn’t mean he or she will understand how to augment the model of a star schema or build a report that will perform optimally.
- Conduct due diligence when selecting your BI / DW vendors. Companies often select a vendor because of perceived “less risk” when in reality, it’s the opposite. From executives to business analysts, all stakeholders need to have structured evaluation criteria from both a technology and a business perspective, especially in this economy where every dollar spent needs to be justified. I met with a major medical manufacturer who purchased a so called pre-built product. After numerous meetings with us, the company finally confessed the warehouse that they called ‘custom’ was actually an out of the box offering from a major vendor. They bought it as part of a larger deal not fully understanding what they were actually getting, nor what it would take to get it up and running.
- Obtain the entire total cost of ownership (TCO) for the project up front which includes conducting due diligence with outside resources. I met with a major retail chain that had to fire three system integrators before settling on one that met the minimum expectations to get their ERP as well as their BI layer implemented. I often see companies that will buy what they think is a great deal from a software vendor only to be left with a combination of poor resources and / or sticker shock when they turn to system integrators for implementation quotes. The old adage of “If it sounds too good to be true, it probably is” still applies today.
What are some of the major trends you see in BI / Data Warehousing?
- Organizations are now realizing they need multiple BI tools as no one BI tool addresses the requirements of the enterprise. A lot of corporations who have tried to implement one tool for the masses are now realizing that due to performance issues or various business requirements, they are back on the market for a second or third BI product to fill the gaps. And in some situations, a department, division, or business unit will buy another BI tool out of end user frustration and IT inherits supporting the product by default.
- Organizations with custom data warehouses are taking a second look at pre-built solutions. Overhead for IT has become the 800 pound gorilla. Those custom data warehouses that were built back in the early 2000’s with a team of 20 to 30 people have now become a huge cost center in corporations and their visibility is being heavily scrutinized. Are they delivering business value that exceeds their costs? Is there a pre-built solution in the market today that can drive a lower TCO and more value? Very likely, yes.
- And on that same theme, companies now confess that their business drivers aren’t that unique. Every company used to think they were different in their business drivers and therefore needed a custom data warehouse. But key metrics such as: global vendor spend, single view of a customer, and accurately and timely closing of the ledger are all quite similar across all major industries.
We understand you like to travel. What are some of your favorite vacation spots?
Well, I have a bucket list of sorts. Checked off the list so far are Paris, London, Costa Rica, Hawaii, the French Riviera, and the Greek Isles. Some of my planned trips include Bora Bora, Thailand, and Australia.